On 16 July 2013 13:33, Paolo Bonzini <pbonz...@redhat.com> wrote: > Il 16/07/2013 14:28, Jan Kiszka ha scritto: >> This changed the value read from unassigned memory from -1 to 0. Any >> particular reason or an unintentional change? > > Cut-and-paste (unassigned RAM used to return 0, invalid MMIO used to > return -1, unifying the paths dropped the difference). > > I guess unassigned RAM can read as -1 just fine, so we can just change > unassigned_mem_read to return -1.
Behaviour for accesses to unassigned addresses should really be both bus and CPU specific... thanks -- PMM
